During a recent Solutions QA test (https://solutions.qa.canonical.com/#/qa/testRun/e2a62353-c9aa-4c1e-920d-fd9d45480a65) one of the kubernetes-worker unit install hooks failed to snap install kube-proxy.
2019-07-23 01:07:30 DEBUG install subprocess.CalledProcessError: Command '['snap', 'install', '--channel=1.15/stable', '--classic', 'kube-proxy']' returned non-zero exit status 1.
2019-07-23 01:07:30 ERROR juju.worker.uniter.operation runhook.go:132 hook "install" failed: exit status 1
2019-07-23 01:07:30 DEBUG juju.machinelock machinelock.go:180 machine lock released for uniter (run install hook)
2019-07-23 01:07:30 DEBUG juju.worker.uniter.operation executor.go:74 lock released
2019-07-23 01:07:30 INFO juju.worker.uniter resolver.go:126 awaiting error resolution for "install" hook
2019-07-23 01:07:30 DEBUG juju.worker.uniter agent.go:20 [AGENT-STATUS] error: hook failed: "install"
Is this environment still available and could you run `sudo snap install --channel 1.15/stable --classic kube-proxy` on one of the worker units and report the results?
Unfortunately I can't find the actual reason that the snap install failed anywhere in the crashdump logs. It looks like the snap layer swallows the error. I wonder if it might have been a snap store hiccup since it happened on more than one worker unit.